Text backdoor detection using an interpretable RNN abstract model

Text backdoor detection using an interpretable RNN abstract model

Deep neural networks (DNNs) are known to be inherently vulnerable to malicious attacks such as the adversarial attack and the backdoor attack. The former is crafted by adding small perturbations to benign inputs so as to fool a DNN. The latter generally embeds a hidden pattern in a DNN by poisoning...

Saved in:

書目詳細資料
Main Authors:	FAN, Ming, SI, Ziliang, XIE, Xiaofei, LIU, Yang, LIU, Ting
格式:	text
語言:	English
出版:	Institutional Knowledge at Singapore Management University 2021
主題:	Training Recurrent neural networks Task analysis Motion pictures Data models Analytical models Sentiment analysis Text backdoor detection RNN model abstraction interpretation OS and Networks Software Engineering
在線閱讀:	https://ink.library.smu.edu.sg/sis_research/7118
標簽:	添加標簽沒有標簽, 成為第一個標記此記錄!
機構:	Singapore Management University
語言:	English

相似書籍

Towards interpreting recurrent neural networks through probabilistic abstraction
由: DONG, Guoliang, et al.
出版: (2020)

Stealthy backdoor attack for code models
由: YANG, Zhou, et al.
出版: (2024)

DeepStellar: Model-based quantitative analysis of stateful deep learning systems
由: DU, Xiaoning, et al.
出版: (2019)

Evaluation of backdoor attacks and defenses to deep neural networks
由: Ooi, Ying Xuan
出版: (2024)

Deepcause: Verifying neural networks with abstraction refinement
由: NGUYEN HUA GIA PHUC,
出版: (2022)

BADFL: Backdoor attack defense in federated learning from local model perspective
由: ZHANG, Haiyan, et al.
出版: (2024)

Scalable and precise refinement of cache timing analysis via path-sensitive verification
由: Chattopadhyay, S., et al.
出版: (2014)

Clean-label backdoor attack and defense: an examination of language model vulnerability
由: Zhao, Shuai, et al.
出版: (2025)

Which neural network makes more explainable decisions? An approach towards measuring explainability
由: ZHANG, Mengdi, et al.
出版: (2022)

SampDetox : Black-box backdoor defense via perturbation-based sample detoxification
由: YANG, Yanxin, et al.
出版: (2024)

Privacy-enhancing and robust backdoor defense for federated learning on heterogeneous data
由: CHEN, Zekai, et al.
出版: (2024)

On-device deep multi-task inference via multi-task zipping
由: HE, Xiaoxi, et al.
出版: (2023)

Neuron sensitivity guided test case selection
由: HUANG, Dong, et al.
出版: (2024)

Multi-Abstraction Concern Localization
由: DUY, Tien-Duy B., et al.
出版: (2013)

Fusing multi-abstraction vector space models for concern localization
由: ZHANG, Yun, et al.
出版: (2018)

NPC: Neuron path coverage via characterizing decision logic of deep neural networks
由: XIE, Xiaofei, et al.
出版: (2022)

EXTENDING TOPIC MODELS FOR TEXT ANALYSIS OF CORPORATE RISK DISCLOSURES
由: BAO YANG
出版: (2014)

Efficient and secure federated learning against backdoor attacks
由: MIAO, Yinbin, et al.
出版: (2024)

Neural network semantic backdoor detection and mitigation: A causality-based approach
由: SUN, Bing, et al.
出版: (2024)

Causality analysis for neural network security
由: SUN, Bing
出版: (2024)

Inferring links between concerns and methods with multi-abstraction vector space model
由: ZHANG, Yun, et al.
出版: (2016)

Adversarial learning for coordinate regression through k-layer penetrating representation
由: JIANG, Mengxi, et al.
出版: (2024)

Design and analysis of a general recurrent neural network model for time-varying matrix inversion
由: Zhang, Y., et al.
出版: (2014)

Hyperbolic graph topic modeling network with continuously updated topic tree
由: ZHANG, Ce, et al.
出版: (2023)

Linkbreaker: Breaking the backdoor-trigger link in DNNs via neurons consistency check
由: CHEN, Zhenzhu, et al.
出版: (2022)

Evaluation of backdoor attacks on deep neural networks
由: Mohamed Nur Hazim Bin Mohamed Ghazali
出版: (2025)

Towards an effective and interpretable refinement approach for DNN verification
由: LI, Jiaying, et al.
出版: (2023)

Mitigating backdoor attacks in large language model-based recommendation systems: a defense and unlearning approach
由: Salimin, Joanne Christina
出版: (2025)

A core task abstraction approach to hierarchical reinforcement learning [Extended abstract]
由: LI, Zhuoru, et al.
出版: (2016)

Mental Visual Indexing: Towards Fast Video Browsing
由: Richang Hong, et al.
出版: (2020)

An Intention-Aware Interactive System for Mobile Video Browsing
由: Jun He, et al.
出版: (2020)

User-database interaction at the knowledge level of abstraction
由: Chan, H.C., et al.
出版: (2014)

TLV: Abstraction through testing, learning, and validation
由: SUN, Jun, et al.
出版: (2015)

CARF-net : CNN attention and RNN fusion network for video-based person reidentification
由: Prasad, Dilip Kumar, et al.
出版: (2019)

Neighbor-anchoring adversarial graph neural networks (extended abstract)
由: LIU, Zemin, et al.
出版: (2022)

Scope-aware data cache analysis for WCET estimation
由: HUYNH BACH KHOA
出版: (2011)

Dynamic decision analysis in medicine: A data-driven approach
由: Cao, C., et al.
出版: (2014)

BadSFL: backdoor attack in scaffold federated learning
由: Zhang, Xuanye
出版: (2024)

Holistic combination of structural and textual code information for context based API recommendation
由: CHEN, Chi, et al.
出版: (2022)

Document graph representation learning
由: ZHANG, Ce
出版: (2023)